|Rates are from|
We studied as many of the questions in the exams that we could and decided to aim for a 20 sec target for the NVR questions and 35 secs target for VR questions. We chose this for two reasons:
• in the exam itself there is usually an amount of instruction that the student needs to read first,
• In the exam the student needs longer to tick the (correct) box in paper based tests.
‘How good you are’ at these tests is shown by a number of things. Obviously, the percentage correct is one of them, but the time taken is also very relevant and must be taken into account. Also, the level of difficulty and even the difficulty level of the questions answered versus those not answered to prevent ‘cherry picking’ the easy questions!
For instance: is a score of 95% accuracy at level 6 in 10mins for a 20 question test, better than 85% accuracy at level 10 in 15mins for a 25 question test? The Standard Score aims to address that question.
We developed the Standard Score to allow pupils to gauge their ability especially as the difficulty level changes and the demands of accuracy and speed are balanced. We reviewed a very large number of tests and worked on an algorithm that combined all these factors. We then fixed a score of 100 at a level that we found students looking to pass 11+; pre-tests; or try for scholarships would be likely to attain.
For instance, at level 10 at 20 secs (for NVR) or 35secs (for VR) per question with 100% correct you would get 100. Getting wrong marks decreases the Standard Score, as does taking extra time.
Just out of interest approximately one sixth of all tests taken (approx 5 million questions over 470,000 tests to date) score over 100 Standard Score.
In order to improve the accuracy, a Standard Score can only be calculated when at least 5 questions have been answered in a test.